The OpenAI math thing is still the biggest story now. I want to wrap this up. On the 6th of October, Scott dropped his rebuttal, “An Open Letter To Steven Pinker On AI.” The article initially went viral, with the general online consensus agreeing with Scott and accusing Pinker of being out of touch. But 12 hours later, OpenAI dropped its math proofs, completely eclipsing Scott’s reply. So OpenAI bailed Pinker out by sparing him a response. Meanwhile, Scott’s opportunity to lay out his case was sidetracked. This goes to show how much the media and bloggers are at the mercy of the AI- and Trump-driven media cycle. Sudden, unexpected developments in either case can set fire to a story that was weeks or months in planning.
This happened in 2024 when blogger and podcaster Tracing Woodgrains published a massive exposé alleging that Wikipedia editors and founders abused their power and reputation to “launder” lies. The story dropped on July 10th, 2024, a Wednesday. It had gained considerable momentum by Friday, and it seemed maybe Wikipedia would have to answer for its deeds. Then of course on the 13th, as we know, Trump was nearly fatally shot at a rally, and that pretty much buried the Wikipedia story.
I eventually plan to respond to Scott’s article, in which I side with Pinker, but I want to wrap this math stuff up first. I had no idea the public, if this is defined as a clique of people I follow online, cared so much about the future of math. It was certainly not this way from 2016-2023 or so. Or ever, really. The assumption has always been that math was immune to developments in the outside world. It was the humanities that bled from wokeness and eroding public support, not math. Coders, maybe, with AI replacing them. But never math.
I saw this tweet going viral as a “dunk” on Terence Tao:
This is a slide from Terence Tao's 'Math 2.0' public lecture yesterday evening at Caltech. I am highly confident that, if the cure actually worked, almost any Stage 4 cancer patient would tell you that human comprehension of its mechanism is completely irrelevant to them. pic.twitter.com/LbKeaObeLz
— Andrew Curran (@AndrewCurran_) October 10, 2026
A lot of people are getting this wrong, although Tao’s example is poorly chosen. A better one would have been air travel: no one agrees on precisely how or why airplanes generate lift, but it’s fortunate that they do, and our incomplete knowledge shouldn’t preclude us from flying. That would have been a better analogy. Tao’s quote can be interpreted in many ways, and the kneejerk response online is an example again of the needed nuance lost in this debate when drying to put people into convenient boxes.
Yes, the mechanism of many drugs is poorly understood. But what gets overlooked is that he isn’t saying you can’t take drugs. He’s saying that ideally the drug would work and you’d also understand how and why it works.
Zvi, in a huge response post, writes,
It turns out this was not front page news. Most people did not hear about it. It should have been, but no one in the news business knows what it means, or thinks people would care.
It was all over the news online. I had to make a deliberate effort to avoid seeing tweets or headlines about it. Yeah, if he’s talking about the print edition of the New York Times, there is going to be a delay of 2-3 days for anything that is not topical breaking news (e.g. developments regarding Iran). Also, time zones add further to the delay. If the OpenAI story dropped at 5 pm PDT on October 6th, the NYT would need 12 or so hours to gather information, so we would expect the story to show up as late as the 8th or 9th.
I couldn’t find a Times story on the October release, but that doesn’t prove they skipped it. The Wall Street Journal covered it right away, with a piece about OpenAI publishing hundreds of proofs on Tuesday night. And the Times did report on OpenAI’s earlier math claim, the Erdős conjecture disproof in May 2026.
In contrasting math discoveries with cancer discoveries, Steve Hsu observes:
steve hsu: If your life’s work is a truly important problem – like curing cancer or fusion energy or discovering the true nature of quantum reality – then suddenly getting a solution from AI would be cause for great celebration.
You might ponder for a moment the second order impact on your profession, but that would be overwhelmed by JOY for the gift that you and the rest of humanity have received.
So why sadness instead of joy? Math operates off of credit: names are inseparably attached to results more so than other areas of research. “Curing cancer” is much more of a team effort compared to proving or disproving some long-standing conjecture. Physics labs are huge collaborative processes, compared to solitary math endeavors. There are fewer opportunities for lone geniuses to stand out in other areas of research, due to costs and other factors (good luck building a home physics lab). Turning math into just another collaborative or lab-driven field means that the last holdout for lone geniuses goes away.
Zvi says, “The Mathematicians Are Not Okay.” I somewhat disagree. Contrary to the media and twitter narrative of mathematicians being either livid or scared of AI, the majority are more sanguine. The technology is disruptive, but it’s premature and unsupported by the evidence to say the profession will be obsolete. This nuance is lost in the debate.
He also says, “Mathematics is facing a real problem here. If the AIs prove all the theorems, then our current methods of getting good at understanding math stop working. The traditional way you understand problems is by working to solve them and a solution is often not worth so much if no one understands it.”
The reality, again, is more nuanced than the media hype that “every problem is a stone’s throw away from being solved with AI.” With the exception of a very tiny and shrinking subset of open problems, you are not going to make meaningful math discoveries by just naively prompting, and this includes frontier models on max settings. Unless you know where to look, you will just be burning your money and tokens.
Also, to reiterate: as someone who “does math,” I can tell you mathematicians aren’t exactly the most open-minded people out there. They have their own cliques and fads, like any other profession. So I don’t think they should be immune to criticism, and I don’t think it’s wrong to poke fun at their largely imagined misfortune. It’s mostly the media and Twitter pushing the narrative that “math is in crisis,” and that’s not representative of everyone. Other professions are routinely mocked and ridiculed (there are entire categories of jokes targeted towards lawyers), so I don’t see why math should be an exception.
If math as a profession somehow “becomes obsolete,” this doesn’t mean it must be saved, much as the rise of automobiles didn’t necessitate a bailout of horse and buggy manufacturers. Society collectively decided the loss of some jobs was an acceptable price of economic progress. Mathematicians, being high IQ, should have little trouble finding other employment (again this is assuming the extremely improbable worst-case scenario). Math can’t just “go away,” but how it’s done will change, if it hasn’t already.
Moreover, to reiterate my earlier post on this matter: I support the use of AI in math. But the controversy also reveals how broken the incentive structure of higher ed and academia, math included, really is. So I can sympathize with people using AI. Academics are under intense pressure to produce original research, and much of it goes unread by reviewers and referees, often because it lacks sufficient novelty. If reviewers can’t be bothered to read the papers, why should authors spend time writing them?
AI solves all of this at once: it speeds up the creation of unread papers and also helps with the “novelty aspect.” If AI leads to the destruction of “prestige mathematics” (which, as I said above, it won’t), I don’t see it as a “net loss” for society. Math will continue, but in a more decentralized manner, with amateurs contributing through a network of social media platforms and preprint repositories. It won’t be enough to only search journals or arXiv for relevant literature: Zenodo, Reddit/Twitter, Academia.edu, GitHub, and various lesser-known places will be fair game.
AI is the first tool that, for once, tips the balance of power in favor of “the little guy.” It’s hard to summon much sympathy for “esteemed academics” who became successful under this broken system and now oppose AI to effectively pull up the ladder . AI is a much-needed disruption, although, again, I predict nothing will change much. But if there is anything to take away here, contrary to media and twitter hype, these is no dominant narrative on this matter by the mathematicians themselves, who are much more indifferent or divided on the matter. Math is not just going away. Mathematicians are not despondent about AI. It’s not a crisis. We just need to wait and see.